[Bugfix] Fix using -O[0,3] with LLM entrypoint

[mgoin]

When using the -O argument with the LLMEngine in benchmark_throughput.py, it would be passed a pre-parsed CompilationConfig that it would then try to dump to json, which is unneeded. I fleshed out the cases I found for parsing to also deal with integers actually being passed in.

Command and error before this PR:

VLLM_USE_V1=1 VLLM_ENABLE_V1_MULTIPROCESSING=1 python benchmarks/benchmark_throughput.py --model neuralmagic/Meta-Llama-3.1-8B-Instruct-FP8 --num-prompts 4000 --input-len 128 --output-len 128 --max-model-len 4096 -O 0
Namespace(backend='vllm', dataset=None, input_len=128, output_len=128, n=1, num_prompts=4000, hf_max_batch_size=None, output_json=None, async_engine=False, disable_frontend_multiprocessing=False, model='neuralmagic/Meta-Llama-3.1-8B-Instruct-FP8', task='auto', tokenizer='neuralmagic/Meta-Llama-3.1-8B-Instruct-FP8', skip_tokenizer_init=False, revision=None, code_revision=None, tokenizer_revision=None, tokenizer_mode='auto', trust_remote_code=False, allowed_local_media_path=None, download_dir=None, load_format='auto', config_format=<ConfigFormat.AUTO: 'auto'>, dtype='auto', kv_cache_dtype='auto', quantization_param_path=None, max_model_len=4096, guided_decoding_backend='outlines', distributed_executor_backend=None, worker_use_ray=False, pipeline_parallel_size=1, tensor_parallel_size=1, max_parallel_loading_workers=None, ray_workers_use_nsight=False, block_size=16, enable_prefix_caching=False, disable_sliding_window=False, use_v2_block_manager=False, num_lookahead_slots=0, seed=0, swap_space=4, cpu_offload_gb=0, gpu_memory_utilization=0.9, num_gpu_blocks_override=None, max_num_batched_tokens=None, max_num_seqs=256, max_logprobs=20, disable_log_stats=False, quantization=None, rope_scaling=None, rope_theta=None, hf_overrides=None, enforce_eager=False, max_seq_len_to_capture=8192, disable_custom_all_reduce=False, tokenizer_pool_size=0, tokenizer_pool_type='ray', tokenizer_pool_extra_config=None, limit_mm_per_prompt=None, mm_processor_kwargs=None, enable_lora=False, enable_lora_bias=False, max_loras=1, max_lora_rank=16, lora_extra_vocab_size=256, lora_dtype='auto', long_lora_scaling_factors=None, max_cpu_loras=None, fully_sharded_loras=False, enable_prompt_adapter=False, max_prompt_adapters=1, max_prompt_adapter_token=0, device='auto', num_scheduler_steps=1, multi_step_stream_outputs=True, scheduler_delay_factor=0.0, enable_chunked_prefill=None, speculative_model=None, speculative_model_quantization=None, num_speculative_tokens=None, speculative_disable_mqa_scorer=False, speculative_draft_tensor_parallel_size=None, speculative_max_model_len=None, speculative_disable_by_batch_size=None, ngram_prompt_lookup_max=None, ngram_prompt_lookup_min=None, spec_decoding_acceptance_method='rejection_sampler', typical_acceptance_sampler_posterior_threshold=None, typical_acceptance_sampler_posterior_alpha=None, disable_logprobs_during_spec_decoding=None, model_loader_extra_config=None, ignore_patterns=[], preemption_mode=None, served_model_name=None, qlora_adapter_name_or_path=None, otlp_traces_endpoint=None, collect_detailed_traces=None, disable_async_output_proc=False, scheduling_policy='fcfs', override_neuron_config=None, override_pooler_config=None, compilation_config=CompilationConfig(level=0, backend='', custom_ops=[], splitting_ops=['vllm.unified_attention', 'vllm.unified_v1_flash_attention'], use_inductor=True, inductor_specialize_for_cudagraph_no_more_than=None, inductor_compile_sizes={}, inductor_compile_config={}, inductor_passes={}, use_cudagraph=False, cudagraph_num_of_warmups=0, cudagraph_capture_sizes=None, cudagraph_copy_inputs=False, pass_config=PassConfig(dump_graph_stages=[], dump_graph_dir=PosixPath('.'), enable_fusion=True, enable_reshape=True), compile_sizes=<function PrivateAttr at 0x7b1bdc699ee0>, capture_sizes=<function PrivateAttr at 0x7b1bdc699ee0>, enabled_custom_ops=Counter(), disabled_custom_ops=Counter(), static_forward_context={}), worker_cls='auto', disable_log_requests=False)
INFO 11-26 16:33:47 __init__.py:42] No plugins found.
Traceback (most recent call last):
  File "/home/mgoin/code/vllm/benchmarks/benchmark_throughput.py", line 442, in <module>
    main(args)
  File "/home/mgoin/code/vllm/benchmarks/benchmark_throughput.py", line 329, in main
    elapsed_time = run_vllm(requests, args.n,
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mgoin/code/vllm/benchmarks/benchmark_throughput.py", line 132, in run_vllm
    llm = LLM(**dataclasses.asdict(engine_args))
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mgoin/code/vllm/vllm/utils.py", line 1052, in inner
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "/home/mgoin/code/vllm/vllm/entrypoints/llm.py", line 189, in __init__
    json.dumps(compilation_config))
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mgoin/.local/share/uv/python/cpython-3.12.4-linux-x86_64-gnu/lib/python3.12/json/__init__.py", line 231, in dumps
    return _default_encoder.encode(obj)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mgoin/.local/share/uv/python/cpython-3.12.4-linux-x86_64-gnu/lib/python3.12/json/encoder.py", line 200, in encode
    chunks = self.iterencode(o, _one_shot=True)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mgoin/.local/share/uv/python/cpython-3.12.4-linux-x86_64-gnu/lib/python3.12/json/encoder.py", line 258, in iterencode
    return _iterencode(o, 0)
           ^^^^^^^^^^^^^^^^^
  File "/home/mgoin/.local/share/uv/python/cpython-3.12.4-linux-x86_64-gnu/lib/python3.12/json/encoder.py", line 180, in default
    raise TypeError(f'Object of type {o.__class__.__name__} '
TypeError: Object of type CompilationConfig is not JSON serializable

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can do one of these:

• Add ready label to the PR
• Enable auto-merge.

🚀

[youkaichao]

thanks for the fix!